Your request is recorded. We will email the prioritized findings within one business day.
8 checks, run in the order they appear in the report. Each one names the finding and what it costs you — not a score.
How fast your pages load on a real phone on a real connection — and which specific asset is holding them back.
Largest Contentful Paint, Interaction to Next Paint, and Cumulative Layout Shift, measured against Google's actual thresholds and traced back to the specific asset causing each failure. This half matters twice over now: a slow, badly structured page is also harder for an AI crawler to parse and cite, so speed problems quietly cost you on both sides of the report.
Over 70% of local searches happen on a phone. Tap targets, readability, and whether your call button is where a thumb can reach it.
We load your site on a real handset viewport and try to do what a customer would: find the phone number, read the prices, submit the form. Tap targets under 44px, body text under 16px, and forms that zoom on focus are the usual culprits. Automated scores miss these because a crawler never has to tap anything.
Title tags, meta descriptions, heading structure, and whether Google can crawl and index the pages that matter to you.
Unique titles under 60 characters, descriptions under 155, one H1 per page, and a heading order that does not skip levels — then we verify your important pages are genuinely indexed and not blocked by robots.txt, a stray noindex, or a mispointed canonical. Duplicate service-page titles are the most common finding here and among the fastest to fix.
Real customer questions put to ChatGPT, AI Overviews, Perplexity, and Gemini — and whether your name comes up at all.
We put the questions your customers actually ask — best plumber near me, who should I call for a leaking roof in Costa Mesa — to ChatGPT, Google AI Overviews, Perplexity, and Gemini, and record what comes back. Usually one of three things: you are recommended, a competitor is recommended, or the assistant describes you using details that are two years out of date. Each needs a different fix.
Hours, address, phone, services. AI assembles these from across the web, so one stale listing can quietly misinform everyone.
AI assembles its answer about you from your site, your Google Business Profile, Yelp, directories, and whatever else it has crawled. When those disagree, the model picks one — often not the current one. We check hours, address, phone, and service list across every surface we can find and flag exactly which stale listing is feeding the wrong answer.
Whether your business, services, hours, and reviews are marked up so both search engines and AI can read them.
LocalBusiness, Service, FAQPage, and BreadcrumbList schema validated against Schema.org. This is the layer that lets Google show hours and ratings directly in results, and gives AI systems machine-readable facts rather than guesses. We flag what is missing, malformed, or contradicts what your pages visibly say — contradictory schema is worse than none.
Whether search and AI crawlers can actually reach your content, or whether robots.txt and rendering are shutting them out.
Whether GPTBot, PerplexityBot, ClaudeBot, and Google-Extended can actually reach your content, and whether your key pages render without JavaScript. Plenty of businesses block AI crawlers by accident through a copied robots.txt, then wonder why assistants never mention them. This is a five-minute check that occasionally explains everything.
Who is outranking you on Google, who AI recommends instead, and what makes them the easier business to cite.
The two or three businesses ranking above you on Google, and the ones AI recommends instead — often not the same list. We identify the concrete gaps: review counts, content depth, location coverage, citation consistency. Being cited by an AI assistant tends to follow the same signals as ranking, so the two halves usually point at one shared root cause.
We look across the website, search results, local profiles, and AI answers together. A page can look polished and still lose customers in the handoff between those systems.
Every image represents a part of the review a human checks and explains in plain language.
Your business name, website, and where to send the findings. Takes about twenty seconds.
Ram goes through your site against the technical checklist, then runs your customers’ real questions past the AI assistants. All by hand.
Within one business day: both halves in a single email, ranked by what is actually worth fixing first. Yours to keep either way.
No score out of 100, no 40-page PDF you'll never read. You get one ordered list covering both halves — what is costing you customers on Google, what is costing you on AI, and which fixes overlap. These are real examples of the kind of finding that comes back, anonymised but typical.
Six of your nine service pages share the identical title tag: "Services | Your Business Name."
The fixGoogle cannot tell which page answers which query, so it picks the wrong one or none at all. Unique, service-specific titles are the highest-leverage hour of work on the site.
Asked "who does emergency water damage repair in your city", ChatGPT and Perplexity both recommend three competitors. You are not mentioned in either answer.
The fixAll three have consistent hours across Google, Yelp and their own site, plus service pages that name the city explicitly. Yours does not. That consistency is what makes a business easy to cite.
Google AI Overviews describes you as offering a service you stopped providing in 2023, and lists a suite number you moved out of.
The fixThe stale details trace to two directory listings and an old Yelp field. Correcting the source listings is what changes the answer — you cannot edit the model, only what it reads.
Your robots.txt blocks GPTBot and Google-Extended, almost certainly copied in from a template.
The fixYou have been invisible to two of the four assistants this entire time. Removing two lines restores access. This is the single most common self-inflicted GEO problem we find.
Your homepage hero is a 2.4MB PNG. LCP is 4.6s on mobile, and the page is heavy enough that AI crawlers time out before reaching your service list.
The fixOne properly sized WebP fixes the ranking signal and the citation problem at once. This is the pattern that makes reviewing both halves together worth it — one root cause, two symptoms.
Your desktop PageSpeed score is 78 rather than 90+.
Why notNearly all your traffic is mobile and your mobile numbers are fine. Chasing this costs real money for no measurable return — we will tell you when something is not worth fixing.
Every audit is different, which is the point. A five-page site for a painter has different problems than a forty-page dental practice site. We don't run a template — Ram looks at your site and your AI answers by hand, and you get what he finds.
Both halves of search, in one review. The technical half looks at your website: page speed and Core Web Vitals, mobile experience, title tags and meta descriptions, heading structure, indexability, structured data, and internal linking. The AI half looks at how ChatGPT, Google AI Overviews, Perplexity, and Gemini describe your business when someone asks them for a recommendation. You get one consolidated set of findings by email, and a call if you want to walk through it.
It's free, no card and no obligation. We do it because showing an owner where their site is losing customers — and where AI is quietly sending those customers to a competitor — is the most honest sales pitch we have. If you want help fixing what we find, plans start at $299/mo with no setup fee. If not, the findings are yours to keep and act on however you like.
You'll have it by email within one business day. This is a hand review rather than an instant automated score. That's deliberate on both halves: a generated report can tell you your Largest Contentful Paint is 4.1 seconds without telling you which fix is worth your money, and AI answers vary between runs, so a single snapshot would be misleading.
Because they feed each other. AI assistants read your website to describe your business — a slow, badly structured site is harder for them to parse and cite. And a fast site with inconsistent business details will still get described wrongly. Looking at both at once is the only way to see which half is actually holding you back, rather than guessing.
Both individual audits are still available — the website audit at /free-website-audit/ and the AI search audit at /free-ai-audit. Plenty of owners assume they only need traditional SEO and are surprised when the AI half shows competitors getting recommended instead. The full audit costs you the same twenty seconds to request, so we'd suggest letting the findings tell you which half needs attention.
Technical fixes tend to surface in Search Console within about 28 days, since Google measures field data on a rolling 28-day window. Content and structural changes compound more slowly, usually 60–90 days. AI visibility moves on its own schedule and is less predictable — improving the underlying signals raises your odds of being cited, but nobody controls what a model outputs. Anyone promising you a fixed timeline on that half is guessing.
Most agencies never look at the AI half at all. If yours is on top of both, our review will confirm it and you can show them off. If not, this is the cheapest accountability check available to you. Call Ram at (714) 732-8549 with the findings in hand and he'll tell you straight whether what you're paying for is doing anything.
We'll follow up once to see if you have questions, and you can book a 15-minute call with Ram to go through it. After that the ball is entirely in your court — no drip sequences, no repeated pitching. Reach Ram directly at (714) 732-8549 or contact@orrkumedia.com.
Skip the form — call Ram directly and we'll look at both halves together on the phone
Free consultation · Clear next steps
Your whole digital presence — website, Google, reviews, social & AI search — plated by one team. One price. No à-la-carte surprise invoices.
A real strategist reviews your site — then emails what's costing you rankings, and what to fix first.
Your request is recorded. We'll email the prioritized findings within1 business day. A walkthrough is optional.
